Horton

RE: RE: Re:BUG #18369: logical decoding core on AssertTXNLsnOrder()

Hayato Kuroda (Fujitsu) <kuroda.hayato@fujitsu.com>

View thread
From: "Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
To: 'Amit Kapila' <amit.kapila16@gmail.com>, ocean_li_996 <ocean_li_996@163.com>
Cc: Alexander Lakhin <exclusion@gmail.com>, "pgsql-bugs@lists.postgresql.org" <pgsql-bugs@lists.postgresql.org>, "feichanghong@qq.com" <feichanghong@qq.com>
Date: 2024-03-09T04:50:02Z
Lists: pgsql-bugs

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Fix catalog lookup due to wrong snapshot for subtransactions during decoding.

Attachments

Dear Amit, 

OK, so let's consider the approach. Note that at this stage, another failure [1]
was ignored.

Here are patches for all supported versions. The failure on PG12-PG15 has been
fixed on my env, but you can confirm as well.

The approach is almost same as what initially shared by Haiyang [2]. However,
instead of extending ReorderBufferXidSetCatalogChanges(), a new function
ReorderBufferXidSetCatalogChangesEx() was added.

Note again that there were changes also in ReorderBufferAssignChild() and
ReorderBufferCleanupTXN(). Extended ReorderBufferXidSetCatalogChanges would the
create ReorderBufferTXN not as top, however, these transaction would not be
associated with the top one. So there is a possibility that txn->node is invalid.
IIUC, only ReorderBufferAssignChild() calls ReorderBufferTXNByXid with create = true
and create_as_top = false, and they would be immediately associated in below.

```
	/* add to subtransaction list */
	dlist_push_tail(&txn->subtxns, &subtxn->node);
	txn->nsubtxns++;
```

[1]: https://www.postgresql.org/message-id/TYCPR01MB12077573479C5A2471BDE8065F5232%40TYCPR01MB12077.jpnprd01.prod.outlook.com
[2]: https://www.postgresql.org/message-id/6d0e80d6.c1fc.18deeb8120a.Coremail.ocean_li_996%40163.com

Best Regards,
Hayato Kuroda
FUJITSU LIMITED
https://www.fujitsu.com/